Bernstein Conference: Moody's CEO Reveals Microsoft Copilot Integration Has Landed "Teens" of Major Bank Engagements in Weeks, Flags Private Credit and PE Cycle as Next Big Ratings Catalyst
Bernstein 42nd Annual Strategic Decisions Conference, May 28, 2026 — Moody's President and CEO Rob Fauber makes the case for why the company's data moat becomes more valuable, not less, in an AI world
The Microsoft Copilot Deal: Earlier Than It Looks, Faster Than Expected
The most concrete new disclosure from Fauber at the Bernstein conference is that Moody's Microsoft Copilot integration — announced only weeks ago — has already generated engagements in the "teens" with major financial institutions, with a handful already in active pilot discussions. For investors who have grown accustomed to AI partnership announcements that produce little near-term commercial visibility, the speed of engagement is notable, even if signed contracts and revenue remain ahead.
The commercial structure is a bring-your-own-license model for now. A customer with an existing Moody's license can call Moody's content directly into Copilot answers — drafting a credit memo, running peer analysis, accessing structured finance research — without leaving the Microsoft Teams interface. Fauber explained the logic simply: "At the end of the day, what I think we're offering to customers is access to our connected intelligence. If you want to do it through Teams, you want to do it through Claude, you want to call it into your own AI environment, we're fine with that." The goal at this stage is embeddedness and usage, not consumption-based pricing. Revenue follows once pilots convert to signed enterprise agreements with new IP protections and pricing terms.
The Anthropic and internal-AI-workflow angles are part of the same strategy. Moody's has formed dedicated sales teams — described by Fauber as a "SWAT team" — at nearly every major financial institution to drive dialogue around how banks can consume Moody's credit models, research, and company knowledge graph through third-party AI surfaces including Rogo, Hebbia, Claude, Teams, and OpenAI. The pipeline, by Fauber's account, is active and building.
Why the Data Moat Is Harder to Replicate Than It Appears
The defensibility question came up directly, and Fauber's answer is worth unpacking carefully because investors who focus only on company-level financial data are missing the breadth of what Moody's actually owns. The asset base has four distinct layers, each with different competitive dynamics.
First is Moody's proprietary research — simply unavailable elsewhere. Second is the banking credit franchise: a contributory default database curated over three decades that calibrates Moody's public and private credit models. These models are used in loan origination, bank credit departments, and regulatory reviews. As Fauber put it, "When your regulator comes in and looks at the loan file and they know you're using the Moody's scoring models, they know that is being calibrated against actual default history." That regulatory credentialization is a meaningful switching cost that deregulation alone does not dissolve.
Third is the catastrophe and actuarial model estate in insurance, built from contributed claims data from the insurance industry itself — in some cases with the customer community co-investing in the tooling. Fourth, and most contested, is the Orbis company database. Fauber openly acknowledged that basic company information can be scraped today. "That's not where the value is." The value lies in two things that are difficult for AI agents to replicate: private data requiring active contractual relationships with hundreds of national information providers who are "becoming more conscious of who is consuming data about companies in their country," and the derived ownership hierarchy and entity relationship data built on top. "That is where the value is," he said, specifically flagging financial crime compliance as the primary use case.
Ratings Guidance Held — and the PE Exit Cycle Remains an Uncaptured Upside
Fauber's refusal to cut ratings guidance after the geopolitical shock that closed April is worth flagging. He referenced what he described as an extraordinary statistic: roughly 80% of U.S. investment-grade issuance in March came in just six days. The message is that financing demand remains deep; what changes is the window. His view is that spreads have recovered, the macroeconomic environment has proven more resilient than feared, and hyperscaler issuance — which nearly matched full-year expectations in the first quarter alone — is not finished.
More importantly, Fauber identified the private equity exit and leveraged finance cycle as the most significant remaining upside in ratings that has not yet materialized. "That PE exit and M&A cycle hasn't really kicked into high gear," he said. "When it does, it is a very virtuous commercial cycle for us" — with multiple rating opportunities across M&A, leverage finance, loans securitizing into CLOs, and CLO ratings themselves. That virtuous chain has historically been among the most powerful revenue multipliers in the ratings business, and by Fauber's read, it is still largely ahead of the company.
Private Credit: From Disintermediation Fear to 80% Growth
Private credit, once the subject of significant investor anxiety about ratings disintermediation, grew roughly 80% in the first quarter. Fauber acknowledged Moody's was "a little slow on the draw" in building out methodologies and go-to-market coverage for private credit, but argued the thesis has inverted. The market, he said, now has a "much broader understanding of the benefit of third-party credit assessment" in private markets, even if the form differs from public ratings.
The opportunity is flowing through structured finance — asset-backed finance and fund finance, which Fauber described as a "booming $1 trillion ecosystem" — as well as through demand for Moody's credit scoring and probability-of-default products from investors seeking third-party views of middle-market credit risk without necessarily purchasing a full rating. The competitive environment in structured finance is more crowded than pre-financial crisis, with Fauber describing what was once a 2.5-agency market now functioning more like a six-agency market in the more transactional, plain-vanilla segments. Coverage levels in private credit ebb and flow more than in fundamental ratings, but Moody's relationship-driven position in corporate credit has seen "very little change" in competitive dynamics since the financial crisis.
The New MA CEO and the Growth Problem She Has Been Hired to Solve
The hire of Cristina Kosmowski — previously a founding member of Salesforce's customer success organization, Chief Customer Officer at Slack during its scale from $90 million to $1 billion, and most recently CEO at Vista-backed LogicMonitor — is a deliberate signal about what Fauber believes is holding back Moody's Analytics growth. The business has nearly doubled in revenues over five years, subscription mix has risen from 80% to 95%, and yet ARR growth has been stuck. Fauber's diagnosis is frank: "We have a fairly complex product array, we have had predominantly field sales, and there's a gravity to that selling model when you're at about $4 billion in revenues." The result has been revenue deceleration because a complex product suite without a well-developed partner channel is hard to scale.
Kosmowski's mandate is to simplify product, pricing and packaging, build a partner ecosystem, and reduce the friction for banks to consume more Moody's content. Fauber noted that the business is also integrating 13 different technology stacks into one platform with a single sales force, work that has been ongoing and costly but is contributing to the margin expansion now being realized. The medium-term margin target remains on track; the question is whether a different go-to-market playbook can re-accelerate the top line.
AI-Related Issuance: A Revenue Mix Nuance Investors Should Understand
Fauber provided a useful framework for how AI infrastructure issuance flows through the ratings P&L. Hyperscalers are frequent issuers, and frequent investment-grade issuers receive volume discounts — meaning issuance growth from that cohort runs ahead of revenue growth, which Moody's describes as "revenue mix unfriendly." However, data center build-out is also generating issuance in project finance, CMBS, CLOs, and power and utility ratings — complex asset classes that are "revenue mix friendly," where transaction revenue growth outpaces volume growth. The net effect from AI-related issuance is a blend of both, and Fauber was careful to note that AI CapEx is one of several medium-term funding drivers, not a single thesis the business is banking on.
Margins and the Agentic Efficiency Opportunity
On margins, Fauber cited a specific example of operational AI deployment in the ratings business: automated QA checks that previously required two separate human review teams. Agents now handle roughly a quarter of those checks, producing both time savings and measurable headcount reduction. Fauber's broader framing is that engineering efficiency — from agentic coding to reduced product development cycle times — is generating investment capacity that gets allocated partly to growth investments and partly to margin improvement. At a group level, Moody's is running at 53% adjusted margins, with the ratings agency in the high 60s. The question of how much higher margins can go is largely a function of how aggressively management chooses to reinvest efficiency gains.
M&A Philosophy: Buy the Data, Not the Software
On capital allocation, Fauber articulated a clear M&A framework that reflects how the company thinks about its core asset. Acquisitions are attractive when they bring proprietary content into the intelligence system that can be monetized across multiple customer segments and workflows — sold "many, many times" in Fauber's words. The secondary criterion is explicit: when evaluating anything that looks like workflow software, Moody's will look hard for a proprietary data asset embedded inside that hasn't yet been monetized. "We might buy something not because of the software, but because of the embedded data asset inside of it." Bureau van Dijk and RMS are cited as the clearest historical examples of this thesis executed well.
The Valuation Argument: "Essential Component of a Broader AI Ecosystem"
Fauber's closing pitch to justify a premium valuation rests on a specific claim about where the AI stack is heading. Referencing Jensen Huang's GTC comment that "structured data is the ground truth of AI," Fauber argued that enterprise AI adoption requires connection to trusted, traceable, auditable content — what he calls "decision grade intelligence" — and that Moody's is assembling precisely that. "AI has a trust problem," he said, echoing a thesis he has written about publicly. The conclusion is that the model race is secondary to the data infrastructure race, and that Moody's connected intelligence system — where every rating, model, forecast, and benchmark resolves to a specific entity and its relationships — is "an essential component of a broader AI ecosystem." Whether the market assigns that value is a separate question, but the strategic logic is internally consistent and the early commercial evidence from the Microsoft and Anthropic integrations suggests real demand, even if the revenue conversion timeline remains uncertain.